I don't know if this is the answer to some of my most pressing current problems, but it does seem silly dumping all the data at once into something that can't handle it.
Maybe if we treat the data as a stream(?) we can allow it to be progressively processed(?)
In their models they get up at a 3-order(!!!) of magnitude(!!!) increase in speed when processing large (and largely simulated) datasets up of tens to hundreds of thousands of proteomics samples.
Worth at least thinking about, IMHO.
No comments:
Post a Comment